UMCC_DLSI-(EPS): Paraphrases Detection Based on Semantic Distance

Second Joint Conference on Lexical and Computational Semantics (*SEM), Volume 2: Seventh International Workshop on Semantic Evaluation (SemEval 2013), pages 93–97, Atlanta, Georgia, June 14-15, 2013. c⃝2013 Association for Computational Linguistics UMCC_DLSI-(EPS): Paraphrases Detection Based on Semantic Distance Héctor Dávila, Antonio Fernández Orquín, Alexander Chávez, Yoan Gutiérrez, Armando Collazo, José I. Abreu DI, University of Matanzas Autopista a Varadero km 3 ½ Matanzas, Cuba. {hector.davila, tony, alexander.chavez, yoan.gutierrez, armando.collazo, jose.abreu}@umcc.cu Andrés Montoyo, Rafael Muñoz DLSI, University of Alicante Carretera de San Vicente S/N Alicante, Spain. {montoyo, rafael}@dlsi.ua.es Abstract This paper describes the specifications and results of UMCC_DLSI-(EPS) system, which participated in the first Evaluating Phrasal Semantics of SemEval-2013. Our supervised system uses different kinds of semantic features to train a bagging classifier used to select the correct similarity option. Related to the different features we can highlight the resource WordNet used to extract semantic relations among words and the use of different algorithms to establish semantic similarities. Our system obtains promising results with a precision value around 78% for the English corpus and 71.84% for the Italian corpus. 1 Introduction It is well known finding words similarity, even when it is lexical or semantic can improve entailment recognition and paraphrase identification; and ultimately lead to improvements in a wide range of applications in Natural Language Processing (NLP). Several areas like question answering, query expansion, information retrieval, and many others, depend on phrasal semantics (PS). PS, is concerned with how the meaning of a sentence is composed both from the meaning of the constituent words, and from extra meaning contained within the structural organization of the sentence itself (Dominey, 2005). The aim of SemEval 2013 competition is also discovering similarity, specifically in Evaluating Phrasal Semantics (EPS). The goal of this task is to evaluate how well systems can judge the semantic similarity of a word and a short sequence of words. That is, given a set of pairs of this type; classify it on negative (if the meaning of the word is semantically different to the meaning of the sequence) or positive (if the meaning of the sequence, as a whole, is semantically close to the meaning of the word). Based on this, we developed a system capable to detect if two phrases are semantically close. The rest of this paper, specifically section 2 is a brief Related Work. Section 3 describes the system architecture and our run. Continuing with section 4 we describe the training phase. Following that, section 5 presents the results and discussion for our Machine Learning System. Finally we conclude and propose our future works (Section 6). 2 Related Work There have been many WordNet-based similarity measures, among other highlights the work of researchers like (Budanitsky and Hirst, 2006; Leacock and Chodorow, 1998; Mihalcea et al., 2006; Richardson et al., 1994). On the other hand, WordNet::Similarity1 (Pedersen et al., 2004) has been used by other researchers in an interesting array of domains. WordNet::Similarity implements measures of similarity and relatedness between a pair of concepts (or synsets2) based …

Content truncated to 500 words